Easy AVR USB Keyboard Firmware and Keymapper

[suicidal_orange]

You were right, there's some kind of hardware issue with my Ducky confirmed the easy way with a spare Teensy and a couple of diodes - if I can get a capital T out of that the board should work.  I can, so it does not.

Found what looked to be left shift connected effectively to the wrong side of it's diode so moved that connection but now I don't have shift keys... This does not explain Soarer's magic and I'm no closer to a usable board but just thought I'd update you!

[samwisekoi]

Update/second round of questions.  Comments in-line and questions at the end.  I am using green text because of the massive nesting, not because I am some dang L33T H@X0R.

Good morning!

I'm de-bugging the GH-122, and some of the faults go away when I jump the appropriate pins on the Teensy++.  That makes them hardware problems for me to chase.

A few problems persist even when I jump pins, which means the problem could be hardware or firmware.  There are three problems affecting multiple keys in columns 18-24 (matrix columns 17-23):

Definitely should fix all the known hardware problems first.  Who knows what effects they could have.

Done. Replaced two diodes and one switch, and now everything from Column_1 through Column_17 is 100%.  I also tested all of the diodes and circuits that are involved, and found no shorts or opens.  I also replaced the diode for KP_DIVIDE, even though it tested good.

#1 An extra KP_DIVIDE scan code is sent by matrix positions:
1,17; 2,17; 2,18; 2,19; 2,20; 2,22; 2,23; 6,17.

#2 Various combinations of extra 0,2,5,8 scan codes (note that those are just under the KP_DIVIDE key!) are sent by matrix positions:
2,21; 3,19; 3,20; 3,21; 3,22; 3,23; 4,17; 4,21; 5,17; 5,21; 5,22; 6,17; 6,21

#3 Extra UP, DOWN, and HOME are sent by matrix positions:
2,21; 3,17; 3,18; 3,19

What are 0,2,5,8 scan codes? 

Key press and then release events for 0, KP_2, KP_5, KP_8.  Specifically Linux reads keysym 0x30, 0xffb2, 0xffb5, and 0xffb8. For example, I get all of those in addition to the correct keysym 0xffb7 when I press KP_7.

Also, If I understand correctly, you're saying that when you hold down any of the listed keys, the PC receives two scancodes instead of one.  Is that right? 

That is correct.  Two or more.

If that's the case, there's really no way for the firmware to screw up that badly, so it must actually be detecting both switches.  We have to figure out why.

Agreed on both points.

Are you absolutely sure that your diodes are installed correctly? 

Yes. Visual inspection and testing with multimeter.

Also, are you finding that problems are localized on PORTA?  I don't have anything that uses an AT90USB128, so I haven't tested it.

No. I am not using Port A as far as I know.  B, C, D, E, and F only.  Unless I don't understand the question, which is entirely possible since I had to read Teensy docs just now to even try.  I'm happy to send you a Teensy++ if you like.

I see nothing in the schematic that would explain this, however problems with the PCB or my assembly are certainly possibilities.  I will proceed with hardware trouble-shooting, but if you could take a quick look at the code to make sure none of the issues are hidden in there, I'd greatly appreciate it.

Also, more of an FYI, matrix position 4,16 doesn't physically exist.  The ISO Enter key uses matrix position 3,16.  I'll send you some more detailed docs later today.

I created the config by eyeballing the picture of the PCB.  I'm sure it has bugs in it.  It would be a very good idea for you to proof-read it.

Will do.

It's very frustrating trying to troubleshoot hardware I don't physically have.  A half-hour job turns into a week-long endeavor.  Hopefully between the two of us we can get it working soon.

Again, agreed on both points.

The problems are almost entirely isolated to the top two rows of the Navigation and Keypad cluster, and the vertical columns below PRINT and KP_DIVIDE.  That makes it seem like a hardware problem, although I cannot think of a combination of shorts, opens, and bad diodes that would explain this behavior.

For example a keypress at matrix location 3,20 (KP_7) should do the following:
- Column pin E0 goes high.
- Switch and diode at 3,20 allow row pin D4 to go high,
- A scan of row pin D4 shows high, resulting in keypress and release events for KP_7

But in addition...
- C2/B1 shows high, resulting in keypress and release events for 0 (or ")" if I test while SHIFT is pressed.)
NOTE: I just tested to verify that the map for the keypad has the keypad zero key outputting KP_0.  It does.
- D6/D2 shows high, resulting in keypress and release events for KP_2.
- D6/D3 shows high, resulting in keypress and release events for KP_5.
- D6/D4 shows high, resulting in keypress and release events for KP_8.

That is just one example, and I am trying to work out a combination of bad diodes, shorts, and opens that would result in those scan results.  I accept that the high on E0 could escape through a bad diode, and even that the entire column under KP_DIVIDE is hosed thereby, but how does the 0) key get involved?

I'll bring my test PCB and multimeter home with me in case you have a brilliant flash or a question.  The current .DAT and .HEX files are attached.

Thanks as always!

 - Ron | samwisekoi

[metalliqaz]

The problems are almost entirely isolated to the top two rows of the Navigation and Keypad cluster, and the vertical columns below PRINT and KP_DIVIDE.  That makes it seem like a hardware problem, although I cannot think of a combination of shorts, opens, and bad diodes that would explain this behavior.

For example a keypress at matrix location 3,20 (KP_7) should do the following:
- Column pin E0 goes high.
- Switch and diode at 3,20 allow row pin D4 to go high,
- A scan of row pin D4 shows high, resulting in keypress and release events for KP_7

But in addition...
- C2/B1 shows high, resulting in keypress and release events for 0 (or ")" if I test while SHIFT is pressed.)
NOTE: I just tested to verify that the map for the keypad has the keypad zero key outputting KP_0.  It does.
- D6/D2 shows high, resulting in keypress and release events for KP_2.
- D6/D3 shows high, resulting in keypress and release events for KP_5.
- D6/D4 shows high, resulting in keypress and release events for KP_8.

That is just one example, and I am trying to work out a combination of bad diodes, shorts, and opens that would result in those scan results.  I accept that the high on E0 could escape through a bad diode, and even that the entire column under KP_DIVIDE is hosed thereby, but how does the 0) key get involved?

I'll bring my test PCB and multimeter home with me in case you have a brilliant flash or a question.  The current .DAT and .HEX files are attached.

Thanks as always!

 - Ron | samwisekoi

This is quite vexing.  I can only come up with three ideas.

First, that you have some kind of crazy hardware problem.  However this board is not very complicated at all.  We can assume the Teensy is all good, so I don't think this is the problem.

Second, that you have the strobe_cols and strobe_low configured wrong.  This could leave pins floating which makes it conceivable that nearby rows affect each other through inductive coupling.  However, since the entire left half of your board works correctly, this is probably not the problem.  Still, there are only four possibilities, watch your PMs for some test builds.

NOTE the series of events you described as "pin goes high" does not happen with the gh122 config file currently in Github.  It has strobe_cols = False and strobe_low = True.  That means it selects rows then reads the columns, and it selects the rows by bringing it down to 0, with all non-selected rows set to 1.  Does your hardware still make sense using that logic?  If not, we will have to change it.

Third, that the software is overrunning its schedule and not handling it gracefully.  I had this idea because all your problems are localized to the far end of the board.  For a matrix that large, it's basically guaranteed to overrun the schedule, but it shouldn't lead to problems like this.  I'll review that part of the code, maybe it was damaged in the rewrite.

So thats what I have so far.  Nothing like a good puzzle, right?

One final thing, if 4,16 doesn't exist, what is the correct location of ANSI enter?

[samwisekoi]

Thanks very much. I need to grok your reply before I can make a sensible next step. The schedule overrun feels like a good fit,  but let's see. I will try those builds in the morning.

I think the spreadsheet has the ANSI Enter location. ISO Enter uses the backslash switch. Verification to follow. 

Thanks again,

 - Ron | samwisekoi

[samwisekoi]

OK, typing on the GH-122 once again.  Build FT works great!

Here are some numpad numbers:

/*-+
789
456
123
0.

There may be some timing issues, but I want to get stabs and all keycaps on the board before declaring that to be true.

Also, I'd like to get the FT version of EKM so I can compile a complete map and test that.

Anyhow, FT is a big step forward, and I thank you very much!

 - Ron | samwisekoi

p.s. For GH-122 buyers who are watching this, it looks like we will soon be good to ship.  More info on that in my own thread. If you ARE lurking here, please extend your thanks to metalliqaz, without whom we would not have the firmware we need!

[metalliqaz]

OK, typing on the GH-122 once again.  Build FT works great!

Here are some numpad numbers:

/*-+
789
456
123
0.

There may be some timing issues, but I want to get stabs and all keycaps on the board before declaring that to be true.

Also, I'd like to get the FT version of EKM so I can compile a complete map and test that.

Anyhow, FT is a big step forward, and I thank you very much!

 - Ron | samwisekoi

p.s. For GH-122 buyers who are watching this, it looks like we will soon be good to ship.  More info on that in my own thread. If you ARE lurking here, please extend your thanks to metalliqaz, without whom we would not have the firmware we need!

The FT build is the same one you're already using.  Just make a new layout and immediately build it and you should get the same file.

[MOZ]

What is this FT build?

[metalliqaz]

What is this FT build?

One of four builds I gave him with all possibilities of strobe_cols and strobe_low

[MOZ]

Anyone using the Originative Ori60 willing to try the attached file.

Ori60 requires you to export the firmware as a bin file. Easy AVR recently added this functionality. Just replace the FLASH.BIN with a new FLASH.BIN generated by Easy AVR.

P.S, AFAIK the KC60 also uses the same matrix and the .py file should work for that as well.

[Data]

@metalliqaz: Thanks dude.  You are the god of firmwarez.

[samwisekoi]

The FT build is the same one you're already using.  Just make a new layout and immediately build it and you should get the same file.

Very interesting.  I made a very small build of the GH-122 withou8t assigning any of the extra keys.  It works better, with only the occasional KP_8 creeping in unwanted.

I'll add more key assignments to see iff perhaps we have an OOM problem. (I see a bit of key-bou8nce as well.)

The above was typed with the GH-122.  Keyboard-generated extra keystrokes remain as they happened.

Thanks,

 - Ron | samwisekoi8

p.s. ANSI Enter is at matrix positi8o8n (4,14).  ISO Enter is at matrix positio8n (3,16).

[samwisekoi]

So I populated the entire matrix, using just plain alphas in all of the extra po8sitio8ns.  The extra KP_8 issu8e remains, and the o258258258258ccai258si258o258258258nal extra KP_DIVIDE has been added.  (The lowercase "I" does thi825s.  <changing keyboards>

The lowercase "i" tosses out extra KP_8, KP_2, and KP_5.  It did not do that with a smaller matrix, and now it seems to happen only after typing for a little bit.  Could there be a memory leak?  I ran every switch through XEV and with the exception of an extra KP_8 when pressing KP_7 and a very occasional extra KP_DIVIDE, all of the keys ran clean.  I will now switch back to the GH-122 and attempt to type every printable character here.  (Remember, the upper row and the left bank have the alphabet.)

Code: [Select]

abcdefghijklmno
r`1234567890-=/*-
stqwertyuiop[]\7889+
uvasdfghjkl;'456
wxzxcvbnm,./123
yz0.

<switching keyboards>
So that went well.  No surprises. I think that tells us that the keyboard CAN type with a full single layer, at least in a tap-tap-tap mode.  Now I'll try some typing:

Code: [Select]

The qu8ick brown fox jumped over the lazy dog.  The lazy dog was really ticked off by this and drafted an urgent tweet to zer followers and then updated zer Tumblr.

<switching keyboards>
OK, not bad.  Just the single KP_8 tossed in, but no KP_DIVIDE or i825 strings.  Shall I fill up some layers and see what happens then?  Is there some way for me to monitor memory utilization?  (FYI, I ordered a couple more Teensy++ for testing purposes.  I can send you one, with or without a giant PCB.)

Next steps?

 - Ron | samwisekoi

[metalliqaz]

The FT build is the same one you're already using.  Just make a new layout and immediately build it and you should get the same file.

Very interesting.  I made a very small build of the GH-122 withou8t assigning any of the extra keys.  It works better, with only the occasional KP_8 creeping in unwanted.

I'll add more key assignments to see iff perhaps we have an OOM problem. (I see a bit of key-bou8nce as well.)

The above was typed with the GH-122.  Keyboard-generated extra keystrokes remain as they happened.

Thanks,

 - Ron | samwisekoi8

p.s. ANSI Enter is at matrix positi8o8n (4,14).  ISO Enter is at matrix positio8n (3,16).

Well, this certainly complicates things.  If the mapping affects how it detects keys, then I am going to need some new ideas.

[metalliqaz]

Next steps?

Can you make sure you're running the latest version?  Your "GH-122 All Keys 160420a.hex" build looks nothing like my test builds.

[samwisekoi]

Next steps?

Can you make sure you're running the latest version?  Your "GH-122 All Keys 160420a.hex" build looks nothing like my test builds.

I was not.  I was on 2_01_08; retrying with 2_01_13 now.

[metalliqaz]

Next steps?

Can you make sure you're running the latest version?  Your "GH-122 All Keys 160420a.hex" build looks nothing like my test builds.

I was not.  I was on 2_01_08; retrying with 2_01_13 now.

Actually, can you use this one please?  This has a fixed GH122 layout.  (ANSI and ISO Enter key)

https://www.dropbox.com/s/vruj5qzl48uaovo/easykeymap_windows_2_01_13b.zip?dl=0

[samwisekoi]

Will download and use the new one, but 2.01.13 seemed to work.  I am typing this on the GH-122.

Woo-hoo!  I'll build a more complicated map using 13b, but8 whatever changed since 2.01.08 seems to have made the Teensy++ much happier!

Thanks VERY MUCH!

 - Ron | samwisekoi

[metalliqaz]

but8

Not entirely happy, it would seem...

In other news I just tested my Sigma board for the first time.  Works perfectly.  *sigh*

[samwisekoi]

but8

Not entirely happy, it would seem...

In other news I just tested my Sigma board for the first time.  Works perfectly.  *sigh*

Go Si888gma!  Oh, look.  Lots of extra 8s.  That i8s good lu8ck i8n Chi8nese.

INTERESTINGLY, IT DOESN'T HAPPEN WHEN I HAVE CAPS LOCK ON.  ALSO, FYI, THAT IS A KP_8 BEING TOSSED AFTER THE "i" SO IT IS AT LEAST COMING FRO8M THE SAME ROW.  OH LOOK - ANOTHER 8 AFTER THE O!

iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii <-- That was typed with CAPS plu8s SHIFT.

Speed seems to matter a bu8nch.  Also, this only happens to letters on the same row as KP_8.  And i8mntermi8ttently - becau8se that i8s the best ki8nd of bu8g.

Shall I do a gratuitous replacement of the KP_8 diode?

 - Ron | samwisekoi

[suicidal_orange]

In other news I just tested my Sigma board for the first time.  Works perfectly.  *sigh*

Not sure if that's a sigh of disappointment that you have no tweaking to do or of delight because your shiny new project works - congrats either way

[metalliqaz]

In other news I just tested my Sigma board for the first time.  Works perfectly.  *sigh*

Not sure if that's a sigh of disappointment that you have no tweaking to do or of delight because your shiny new project works - congrats either way

Yeah that came off wrong.  I'm glad my project is working but I'm frustrated that the good bugs are always on boards that I don't have.

[samwisekoi]

In other news I just tested my Sigma board for the first time.  Works perfectly.  *sigh*

Not sure if that's a sigh of disappointment that you have no tweaking to do or of delight because your shiny new project works - congrats either way

Yeah that came off wrong.  I'm glad my project is working but I'm frustrated that the good bugs are always on boards that I don't have.

I got it!

[samwisekoi]

What was the change between 13 and 13b?  More specifically, should I go back to 13?

 - Ron | samwisekoi

[metalliqaz]

What was the change between 13 and 13b?  More specifically, should I go back to 13?

 - Ron | samwisekoi

It fixed the mapping of the Enter keys.  Don't go back to the older version, this is the future, man.

I noticed the code wasn't disabling JTAG for the AT90USB1286 builds.  It's probably not going to help, but please try this newest version (13c):

https://www.dropbox.com/s/1j6xxz94uw74s0v/easykeymap_windows_2_01_13c.zip?dl=0

[SpAmRaY]

Watching all this debugging with great anticipation. Thanks samwisekoi for thoroughly testing and metalliqaz for all the countless hours put into easy avr.

[samwisekoi]

Thanks spam and metal'z .

Trying to get some altitude over this problem, I wonder if we've finally hit the scale where analog circuitry starts to exhibit analog side effects. We treat modern keyboards like digital devices, but really they are not. Does an almost two square foot board need a ground plane, thicker traces, or even inductive shielding? If so, we've entered Haata land.

Alternatively, it may just be JTAG.  )

Just some Thursday night pondering.

 - Ron ¦ samwisekoi

[suicidal_orange]

What happens when you connect a matrix column to two pins?  If all the connections are in place - nothing out of the ordinary.  If you scan the matrix a certain way (Soarer's) - nothing out of the ordinary.  If you scan it the opposite way (EasyAVR) and one switch is only connected to one of the pins that switch misbehaves!

Turns out my shift key's diode connection was not connected to the other diode in it's row (right shift) and as they're nowhere near each other newbie me ran a second wire to the Teensy.  Newbie me wondered why anyone would design a matrix with just two keys in a row/column when there were a couple of candidates with those positions empty and remembers checking extensively before accepting that they just did, but newbie me failed - there was a connection, and both Shift keys should be in column 10 as pictured before.

Sorry for wasting a very small fraction of your time metalliqaz, and thanks for being so insistent it was a hardware problem - I have learned another possible symptom of a matrix short because of it

Going to add a couple more indicator LEDs while I have it open then will test if my non-huge matrix exhibits any strangeness as it's probably the only other EasyAVR'd AT90USB1286 board in existence other than sam's.

[metalliqaz]

Thanks spam and metal'z .

Trying to get some altitude over this problem, I wonder if we've finally hit the scale where analog circuitry starts to exhibit analog side effects. We treat modern keyboards like digital devices, but really they are not. Does an almost two square foot board need a ground plane, thicker traces, or even inductive shielding? If so, we've entered Haata land.

Alternatively, it may just be JTAG.  )

Just some Thursday night pondering.

 - Ron ¦ samwisekoi

To be honest, at 5v and sub-kHz frequencies, we really don't have much to worry about.  If you try to read floating pins then you can expect it to go haywire, but otherwise I expect it to be solid.  If there is an abnormally large load on the pins, then increasing the matrix setup wait time in the config console would take care of it.  You should try that, by the way, in addition to the PORTF fix in 13c.

Quick question.  When you get the phantom KP8 keypress, does it stay "on" as long as you hold the other key, or is it a fleeting scancode that shows up and immediately disappears?

[metalliqaz]

What happens when you connect a matrix column to two pins?  If all the connections are in place - nothing out of the ordinary.  If you scan the matrix a certain way (Soarer's) - nothing out of the ordinary.  If you scan it the opposite way (EasyAVR) and one switch is only connected to one of the pins that switch misbehaves!

Turns out my shift key's diode connection was not connected to the other diode in it's row (right shift) and as they're nowhere near each other newbie me ran a second wire to the Teensy.  Newbie me wondered why anyone would design a matrix with just two keys in a row/column when there were a couple of candidates with those positions empty and remembers checking extensively before accepting that they just did, but newbie me failed - there was a connection, and both Shift keys should be in column 10 as pictured before.

Sorry for wasting a very small fraction of your time metalliqaz, and thanks for being so insistent it was a hardware problem - I have learned another possible symptom of a matrix short because of it

Going to add a couple more indicator LEDs while I have it open then will test if my non-huge matrix exhibits any strangeness as it's probably the only other EasyAVR'd AT90USB1286 board in existence other than sam's.

I'm having a hard time visualizing what you've got over there, but if it is working better I'm thrilled!

[suicidal_orange]

I'm having a hard time visualizing what you've got over there, but if it is working better I'm thrilled!

Have you not seen my wonderous Ducky?  This was my introduction to mech's, I bought it with a couple of broken switches and a break in the matrix and fixed it up.  The matrix is a little bit crazy as you've seen, slightly less so following today's revelation but why they chose to put F5 on a row with two 'shift's, two 'control's and pause/break will forever remain a mystery!  Physically it looks quite like this, but with more spiderweb...

More

First problem found - qwertyuiop][ )


I think this post demonstrates that typing on the primary layer appears to work correctly, the only unused letter seems to be Q and it too work just fine.  I can even do a capital T using just my left hand so a definite improvement!

Still not convinced that the layers are playing nicely but that will check and confirm before I trouble you with it.

Thanks again for your help!

[samwisekoi]

Quick question.  When you get the phantom KP8 keypress, does it stay "on" as long as you hold the other key, or is it a fleeting scancode that shows up and immediately disappears?

OK, re-built the existing .dat file in 13c.  Seems worse.  Here is a snippet:

Code: [Select]

Easy Keymap v2.01.13c/GH-122.2016 160422aE258

This is worse, or so it seems.  Maybe not.

Just on the nav/numpad block, I guess.

257825788888888888888825788888888888888888888888888882578888888888888888888888888888825788888888888888888888888888888

That last bit was to test the run-on KP_8 problem when I kept the KP_7 key held down.  I had not tried that before, but there you go.  Picked up some KP_2 and KP_5 action with this build.

No issues with the main board as far as I can see.

So I did this:

increasing the matrix setup wait time in the config console would take care of it.  You should try that

And settings of 10, 50, 200, and 1, all made the problem worse.

Should I re-do the .dat file from scratch each time your version is updated?  Also, should I replace some diodes for luck?  KP_2, KP_5, and KP_8 come to mind.

Thanks again,

 - Ron | samwisekoi

[FletchINKy]

What level of backlighting support (and modes?) are available on handwired teensy boards?

Looked through the sauce and didn't really grok how it applied to handwired boards... Thanks!

[suicidal_orange]

You were sad to have no bugs so here's one that should be easy to replicate, though probably not as fun to fix.

Assign key 1 as a normal FN
On layer 1 assign key 2 as toggle FN 2
On layer 2 assign key 1 as normal FN3
On layer 3 assign key 2 to a key which you don't mind getting stuck

Hold key 1, press key 2, release both.  Repeat.


I set up QWERTY and Colemak layers with a copy of each to use with an FN (implied Ctrl) and use Insert on the QWERTY copy to go to Colemak and Delete on the Colemak copy for QWERTY, but on pressing Insert in an attempt to switch back my busy terminal was filled with ^[[2~ (Insert) which alerted me to the problem.  Why not just have a "swap" switch?  There will be more layouts/layers, don't want to have to cycle through them all.  The current layout's direct switch will be left as it's default key as simulated above.

[suicidal_orange]

What level of backlighting support (and modes?) are available on handwired teensy boards?

Looked through the sauce and didn't really grok how it applied to handwired boards... Thanks!

What components are you handwiring?  Anything that can be done on a PCB can in theory be handwired, the firmware doesn't know or care

[FletchINKy]

What components are you handwiring?  Anything that can be done on a PCB can in theory be handwired, the firmware doesn't know or care

I guess my question is more related to what Easy AVR supports... easily... 

If I wire a 5v LED strip to one of the pins and indicate it as a backlight correctly, can I just assign a key to SCANCODE_BL_MODE and flip through the... modes? What modes? The docs say the zones are explained elsewhere, but I haven't been able to find that documentation, or do zones not apply to handwired boards?

Forgive my ignorance, just not familiar with how the backlights are programmed.

[metalliqaz]

What components are you handwiring?  Anything that can be done on a PCB can in theory be handwired, the firmware doesn't know or care

I guess my question is more related to what Easy AVR supports... easily... 

If I wire a 5v LED strip to one of the pins and indicate it as a backlight correctly, can I just assign a key to SCANCODE_BL_MODE and flip through the... modes? What modes? The docs say the zones are explained elsewhere, but I haven't been able to find that documentation, or do zones not apply to handwired boards?

Forgive my ignorance, just not familiar with how the backlights are programmed.

It should be documented, but it isn't.  I haven't got around to writing it.  There is a lot to explain, so if you just post your planned hardware, I can give you the correct config.  For backlighting in particular, you can look at the KMAC, Orion, and Sigma config files for an idea of what it does.

[metalliqaz]

You were sad to have no bugs so here's one that should be easy to replicate, though probably not as fun to fix.

Assign key 1 as a normal FN
On layer 1 assign key 2 as toggle FN 2
On layer 2 assign key 1 as normal FN3
On layer 3 assign key 2 to a key which you don't mind getting stuck

Hold key 1, press key 2, release both.  Repeat.


I set up QWERTY and Colemak layers with a copy of each to use with an FN (implied Ctrl) and use Insert on the QWERTY copy to go to Colemak and Delete on the Colemak copy for QWERTY, but on pressing Insert in an attempt to switch back my busy terminal was filled with ^[[2~ (Insert) which alerted me to the problem.  Why not just have a "swap" switch?  There will be more layouts/layers, don't want to have to cycle through them all.  The current layout's direct switch will be left as it's default key as simulated above.

What the hell are you people doing to me?  Switching back and forth between Colemak and Qwerty... are you insane?  I can't even parse out this mad science.

I'll look into it but there already is "swap": it's a FN key in toggle mode.  If you think about it, it's the same thing.

[metalliqaz]

You were sad to have no bugs so here's one that should be easy to replicate, though probably not as fun to fix.

Assign key 1 as a normal FN
On layer 1 assign key 2 as toggle FN 2
On layer 2 assign key 1 as normal FN3
On layer 3 assign key 2 to a key which you don't mind getting stuck

Hold key 1, press key 2, release both.  Repeat.

I just set this up in my Zeta and it works correctly.  There must be something more to your config than this.

[metalliqaz]

Quick question.  When you get the phantom KP8 keypress, does it stay "on" as long as you hold the other key, or is it a fleeting scancode that shows up and immediately disappears?

OK, re-built the existing .dat file in 13c.  Seems worse.  Here is a snippet:

Code: [Select]

Easy Keymap v2.01.13c/GH-122.2016 160422aE258

This is worse, or so it seems.  Maybe not.

Just on the nav/numpad block, I guess.

257825788888888888888825788888888888888888888888888882578888888888888888888888888888825788888888888888888888888888888

That last bit was to test the run-on KP_8 problem when I kept the KP_7 key held down.  I had not tried that before, but there you go.  Picked up some KP_2 and KP_5 action with this build.

No issues with the main board as far as I can see.

So I did this:

increasing the matrix setup wait time in the config console would take care of it.  You should try that

And settings of 10, 50, 200, and 1, all made the problem worse.

Should I re-do the .dat file from scratch each time your version is updated?  Also, should I replace some diodes for luck?  KP_2, KP_5, and KP_8 come to mind.

Thanks again,

 - Ron | samwisekoi

We will get this, we just have to narrow the solution space.

PM'ing two test builds.  Please let me know how they work.  One only implements the normal fullsize keyboard sections of the layout.  If that works then the problem is with the JUMBO build.  The second implements only the problematic right side of the keyboard.  If that works then there is some error in the matrix scan.

[FletchINKy]

It should be documented, but it isn't.  I haven't got around to writing it.  There is a lot to explain, so if you just post your planned hardware, I can give you the correct config.  For backlighting in particular, you can look at the KMAC, Orion, and Sigma config files for an idea of what it does.

Ok I think I'm following. The modes are a list of possible combinations of LEDs including indicators. All 1s is all LEDs active and all 0s is none active. Are on and off the only states supported? While looking through the docs I read about breathing and decay modes, but wondered how you could do those with only a teensy going high or low.

[MOZ]

Ok I think I'm following. The modes are a list of possible combinations of LEDs including indicators. All 1s is all LEDs active and all 0s is none active. Are on and off the only states supported? While looking through the docs I read about breathing and decay modes, but wondered how you could do those with only a teensy going high or low.

The way I think it works is, the modes represents, which "zones" or LEDS to enable. The system checks the tuple consisting of binary values, if it's a
- 0 and an indicator LED, then it's state is governed by it's assignment
- 0 and not an indicator LED, then it is turned off
- 1 and irrespective of indicator or not, it is configured as a backlight LED

Backlight LEDs/zones have different modes, such as on, off, breathing, decay, reactive. This is done using PWM control. Usually people use a transistor as a relay to control the LEDs.

[FletchINKy]

The way I think it works is, the modes represents, which "zones" or LEDS to enable. The system checks the tuple consisting of binary values, if it's a
- 0 and an indicator LED, then it's state is governed by it's assignment
- 0 and not an indicator LED, then it is turned off
- 1 and irrespective of indicator or not, it is configured as a backlight LED

Backlight LEDs/zones have different modes, such as on, off, breathing, decay, reactive. This is done using PWM control. Usually people use a transistor as a relay to control the LEDs.

Just tested with a one LED board, so I wasn't able to test the Indicator vs backlight modes states, but confirmed the backlight on/off, dimmer, and standard modes.

How does the dimmer work? Is it a strobe? Or should I worry about choosing a particular resistor value?

[metalliqaz]

Just tested with a one LED board, so I wasn't able to test the Indicator vs backlight modes states, but confirmed the backlight on/off, dimmer, and standard modes.

How does the dimmer work? Is it a strobe? Or should I worry about choosing a particular resistor value?

MOZ is mostly correct.  The firmware supports multiple modes, zones, and dimmer levels.

Modes are the behavior of the backlight LEDs.  There is On, Off, Breathing, and Reactive modes.
Zones are areas of the keyboard.  For example, my keyboard allows you to turn on the backlights of the alphas, mods, and Function keys separately.
Dimmer levels only take effect for zones that are currently enabled and when mode is set to On.

The array in the config file contains 1 or 0 (true or false) for each LED output, including indicators and backlight pins.  Indicators always act as assigned, but if they are assigned to "backlight", then this array is used to tell the firmware which zone they apply to.  In my Zeta/Sigma, they all part of the mods zone.

Almost everyone should just keep it simple.  All on and all off.  Don't worry about it.

Choose your resistor value for your desired max LED brightness.  Don't worry about the dimming, the firmware will take care of that by pulsing it on and off very quickly.

[FletchINKy]

MOZ is mostly correct.  The firmware supports multiple modes, zones, and dimmer levels.

Modes are the behavior of the backlight LEDs.  There is On, Off, Breathing, and Reactive modes.
Zones are areas of the keyboard.  For example, my keyboard allows you to turn on the backlights of the alphas, mods, and Function keys separately.
Dimmer levels only take effect for zones that are currently enabled and when mode is set to On.

The array in the config file contains 1 or 0 (true or false) for each LED output, including indicators and backlight pins.  Indicators always act as assigned, but if they are assigned to "backlight", then this array is used to tell the firmware which zone they apply to.  In my Zeta/Sigma, they all part of the mods zone.

Almost everyone should just keep it simple.  All on and all off.  Don't worry about it.

Choose your resistor value for your desired max LED brightness.  Don't worry about the dimming, the firmware will take care of that by pulsing it on and off very quickly.

Interesting, so if I were to do an RGB strip I would be able to wire the R, G, and B to separate pins and through zones define color combinations, but not individual pin brightness/pulsing/dimming?

[metalliqaz]

MOZ is mostly correct.  The firmware supports multiple modes, zones, and dimmer levels.

Modes are the behavior of the backlight LEDs.  There is On, Off, Breathing, and Reactive modes.
Zones are areas of the keyboard.  For example, my keyboard allows you to turn on the backlights of the alphas, mods, and Function keys separately.
Dimmer levels only take effect for zones that are currently enabled and when mode is set to On.

The array in the config file contains 1 or 0 (true or false) for each LED output, including indicators and backlight pins.  Indicators always act as assigned, but if they are assigned to "backlight", then this array is used to tell the firmware which zone they apply to.  In my Zeta/Sigma, they all part of the mods zone.

Almost everyone should just keep it simple.  All on and all off.  Don't worry about it.

Choose your resistor value for your desired max LED brightness.  Don't worry about the dimming, the firmware will take care of that by pulsing it on and off very quickly.

Interesting, so if I were to do an RGB strip I would be able to wire the R, G, and B to separate pins and through zones define color combinations, but not individual pin brightness/pulsing/dimming?

That's right.

[TheGlow]

hey metalliqaz, slight update.
I was trying to tweak my debounce and matrix wait times.
So debounce 20, matrix wait 60 still had the doubles and delayed hits.
Worked matrix up to 80, now I'm at 100 and the typos have dropped drastically.
It's been over a week now I believe and Im getting 0-3 typos a day.
Also the letter H and spaces would be the most typoed. I haven't had a single H yet. Space is a bit random still.

Again I am a bit puzzled as it must be tied into the way I type somehow.
At home on my razer blackwidow cherry mx blues I get some of the typos as well. But if so I think its only rare cases like letter a coming one more time later.

[Cadriel]

Finally got my GH60 Satan in-house. Dumped Easy AVR on the board and all is well!

Thanks everyone for the effort to get this PCB working - loving it!

[metalliqaz]

Finally a new feature.  With the next release you will be able to program your board without leaving the keymapper application.
This thanks to MOZ's development for his upcoming project.

[metalliqaz]

It currently supports Teensy loader, Flip, and dfu-programmer.  When the new version is released, some testing on other platforms would be appreciated.

[MOZ]

Yeay, this is exciting. I'll test it out today.